asr_lib 0.6.0

On-device speech recognition for Flutter powered by sherpa-onnx.

example/example.md

sherpa-asr 示例

1. 直接初始化

import 'package:asr_lib/asr_lib.dart';
import 'package:sherpa_onnx/sherpa_onnx.dart' as sherpa_onnx;

final asr = AsrLib();

final modelConfig = sherpa_onnx.OfflineModelConfig(
  senseVoice: sherpa_onnx.OfflineSenseVoiceModelConfig(
    model: '/path/to/senseVoice/model.int8.onnx',
  ),
  tokens: '/path/to/senseVoice/tokens.txt',
  numThreads: 2,
  provider: 'cpu',
);

await asr.initialize(
  config: AsrConfig(modelConfig: modelConfig),
);
await asr.startRecording();

2. 先下载默认模型

await asr.downloadModel();

默认会下载：

• senseVoice/model.int8.onnx
• senseVoice/tokens.txt